label llms.txt open_chatgpt

Seedance 2.5 API Documentation

Generate Seedance 2.5 and Seedance 2.5 Flash videos through API v3: text-to-video, reference images, video and audio, first/last frames, video editing, and 4K output.

API v3 · 2.5 / 2.5 Flash Base URL: https://seedanceapi.org/v3

API v3 is a separate endpoint for Seedance 2.5 only. Existing /v1 and /v2 integrations keep working exactly as before; switch to /v3 only when you want Seedance 2.5.

Quick Start

bash
1curl -X POST 'https://seedanceapi.org/v3/generate' \
2 -H 'Authorization: Bearer YOUR_API_KEY' \
3 -H 'Content-Type: application/json' \
4 -d '{
5 "model": "seedance-2.5",
6 "prompt": "A cinematic aerial shot over a coastline at golden hour",
7 "aspect_ratio": "16:9",
8 "resolution": "720p",
9 "duration": 10
10 }'

Authentication

All API requests require a Bearer token in the Authorization header.

Important: Get your API key from the API Keys page in your dashboard. Keep it on your server and never ship it in browser code. → Get Your API Key

http
1Authorization: Bearer YOUR_API_KEY

Generation modes

ModeFields to send
Text to videoprompt + aspect_ratio
Reference generation (images, video, audio)prompt + image_urls / reference_video_urls / reference_audio_urls
First / last frameprompt + first_frame_url (+ last_frame_url)
Video editingprompt + reference_video_urls (1) + duration: -1 + aspect_ratio: "adaptive"

Pricing (credits)

Credits are charged per second of video and depend on the model, resolution, and whether you send reference videos. The amount is deducted when the task is created and refunded automatically if it fails.

Need to recharge credits? → Top up on the pricing page. 1 USD = 100 credits.

2.5

Seedance 2.5

Highest quality, supports 480p, 720p, and native 4K.

ResolutionInputRateExample
480pText / images / audio28 credits / s10s output = 280 credits
480pWith reference video17 credits / s5s reference + 10s output = 255 credits
720pText / images / audio63 credits / s10s output = 630 credits
720pWith reference video38 credits / s5s reference + 10s output = 570 credits
4kText / images / audio251 credits / s10s output = 2510 credits
4kWith reference video155 credits / s5s reference + 10s output = 2325 credits
Flash

Seedance 2.5 Flash

Faster and cheaper, supports 480p and 720p.

ResolutionInputRateExample
480pText / images / audio18.5 credits / s10s output = 185 credits
480pWith reference video11.5 credits / s5s reference + 10s output = 173 credits
720pText / images / audio41 credits / s10s output = 410 credits
720pWith reference video25.5 credits / s5s reference + 10s output = 383 credits

With reference_video_urls, billing uses the lower video-input rate on reference video seconds plus output seconds. Video editing (duration -1) bills twice the input video length. Totals are rounded up to a whole credit.

API Endpoints

POST/v3/generate

Start a Seedance 2.5 or Seedance 2.5 Flash job. The call returns a task_id right away; poll /v3/status or pass callback_url to get the result.

Request Body

Body ParametersJSON
prompt:string

Text prompt (required, up to 5000 characters).

model:optional string

seedance-2.5 (default) or seedance-2.5-flash. Defaults to seedance-2.5.

duration:optional number

Output length in seconds: any integer from 4 to 30. Use -1 for video editing, where the output keeps the input video's length. Defaults to 5.

resolution:optional string

Seedance 2.5: 480p, 720p, or 4k. Seedance 2.5 Flash: 480p or 720p. 1080p is not available. Defaults to 720p.

aspect_ratio:optional string

16:9, 9:16, 4:3, 3:4, or adaptive. Video editing requires adaptive. Defaults to 16:9.

image_urls:optional string[]

Up to 4 reference image URLs (http/https). Describe how each image should be used in the prompt.

first_frame_url:optional string

Image URL used as the first frame. Cannot be combined with image_urls, reference_video_urls, or reference_audio_urls.

last_frame_url:optional string

Optional image URL used as the last frame. Requires first_frame_url.

reference_video_urls:optional string[]

Up to 3 reference video URLs, 30 seconds in total. For video editing, send exactly one video between 4 and 30 seconds long.

reference_audio_urls:optional string[]

Up to 3 reference audio URLs, for example music or a voice track.

generate_audio:optional boolean

Generate a soundtrack together with the video. Defaults to true.

web_search:optional boolean

Let the model look up the web for subjects in the prompt. Defaults to false.

callback_url:optional string

Optional webhook called when the task finishes (public URL; localhost is not allowed).

public:optional boolean

Optional: allow the result to appear in the public gallery.

Text to video

json
1{
2 "model": "seedance-2.5",
3 "prompt": "A slow dolly shot through a neon-lit street in the rain",
4 "aspect_ratio": "16:9",
5 "resolution": "720p",
6 "duration": 10
7}

Reference image + audio

json
1{
2 "model": "seedance-2.5",
3 "prompt": "The woman in the reference image dances to the beat of the reference audio in a sunlit studio",
4 "image_urls": ["https://example.com/character.jpg"],
5 "reference_audio_urls": ["https://example.com/beat.mp3"],
6 "aspect_ratio": "9:16",
7 "resolution": "720p",
8 "duration": 8
9}

First and last frame (Flash)

json
1{
2 "model": "seedance-2.5-flash",
3 "prompt": "The flower bud slowly opens into full bloom",
4 "first_frame_url": "https://example.com/bud.jpg",
5 "last_frame_url": "https://example.com/bloom.jpg",
6 "resolution": "480p",
7 "duration": 6
8}

Video editing

json
1{
2 "model": "seedance-2.5",
3 "prompt": "Change the season to winter with falling snow; keep the camera motion",
4 "reference_video_urls": ["https://example.com/input-12s.mp4"],
5 "aspect_ratio": "adaptive",
6 "duration": -1,
7 "resolution": "720p"
8}

4K output

json
1{
2 "model": "seedance-2.5",
3 "prompt": "Macro shot of morning dew rolling off a green leaf",
4 "aspect_ratio": "16:9",
5 "resolution": "4k",
6 "duration": 5
7}

Responses

Task created

1{
2 "code": 200,
3 "message": "success",
4 "data": {
5 "task_id": "n90a1b2c3d4e5f6sd25",
6 "status": "IN_PROGRESS",
7 "consumed_credits": 630
8 }
9}
GET/v3/status

Poll by task_id until status is SUCCESS or FAILED. Only tasks created with /v3/generate are returned here.

Query Parameters

Body ParametersJSON
task_id:string

Task id returned by POST /v3/generate (ends with sd25).

Example Request

bash
1curl -X GET 'https://seedanceapi.org/v3/status?task_id=n90a1b2c3d4e5f6sd25' \
2 -H 'Authorization: Bearer YOUR_API_KEY'

Tip: When status is SUCCESS, data.response is an array of video URLs; use data.response[0] for the first URL. Download and store the file if you need to keep it.

javascript
1const videoUrl = data.response[0];

Responses

1{
2 "code": 200,
3 "message": "success",
4 "data": {
5 "task_id": "n90a1b2c3d4e5f6sd25",
6 "status": "SUCCESS",
7 "consumed_credits": 630,
8 "created_at": "2026-11-19T10:30:00Z",
9 "type": "seedance-2.5-t2v",
10 "request": {
11 "model": "seedance-2.5",
12 "prompt": "A slow dolly shot…",
13 "aspect_ratio": "16:9",
14 "duration": 10
15 },
16 "response": [
17 "https://cdn.example.com/videos/n90a1b2c3d4e5f6sd25.mp4"
18 ],
19 "error_message": null
20 }
21}

API Playground

Try the API in the browser. Replace YOUR_API_KEY with your real key. Each successful call is billed.

API PlaygroundPOST

Error Codes

StatusWhen it happens
400 Bad RequestA parameter is missing or invalid, for example an unsupported model, resolution, aspect_ratio, duration, or media URL. The message explains which one.
401 UnauthorizedAPI key missing or invalid.
402 Not enough credits for this task.
404 Not FoundThe task does not exist, or it was not created with /v3/generate.
500 Internal Server ErrorServer error. Retry later; failed tasks are refunded automatically.